nginx: re-resolve backend upstream after container restart #75

New issue

Open

opened 2026-10-26 10:46:09 -04:00 by vmanne-poc · 0 comments

vmanne-poc commented

2026-10-26 10:46:09 -04:00

Owner

Phase 13.1 (v0.15.1) follow-up; related to #29 / #49 API traffic.

nginx resolves the backend upstream hostname once at start, so after the backend container restarts (new IP) API requests fail with 502 until nginx is restarted.

Plan

Use Docker's resolver (resolver 127.0.0.11 valid=10s) with a variable upstream so the name is re-resolved.

Done when

Restarting the backend container alone keeps the API reachable through nginx.

Phase 13.1 (v0.15.1) follow-up; related to #29 / #49 API traffic. nginx resolves the backend upstream hostname once at start, so after the backend container restarts (new IP) API requests fail with 502 until nginx is restarted. **Plan** - Use Docker's resolver (`resolver 127.0.0.11 valid=10s`) with a variable upstream so the name is re-resolved. **Done when** - Restarting the backend container alone keeps the API reachable through nginx.

vmanne-poc added this to the Phase 13.1: API key scopes & provider webhooks milestone

2026-10-26 10:46:09 -04:00

vmanne-poc added the

bug

label

2026-10-26 10:46:09 -04:00

vmanne-poc referenced this issue

2026-10-26 10:46:13 -04:00

Project API keys #29

vmanne-poc referenced this issue

2026-10-26 10:46:14 -04:00

API-based sending providers (SendGrid, Mailgun, Postmark, SES API) #49